Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.bricks.tools/llms.txt

Use this file to discover all available pages before exploring further.

Model Context Protocol (MCP) lets you connect external tool servers to the AI agent. MCP tools extend the agent’s capabilities beyond the built-in tools.

​
Built-in MCP server

BRICKS projects include a built-in bricks-ctor MCP server that is automatically configured during bun install. It provides tools for compiling, previewing, searching icons and Lottie animations, selecting Hugging Face models, and managing media assets.

bricks-ctor built-in MCP server

See the full tool reference in the bricks-ctor docs.

​
Configuration

MCP servers are configured in a .mcp.json file in your project root: 
{
  "mcpServers": {
    "my-server": {
      "command": "npx",
      "args": ["-y", "some-mcp-server"],
      "env": {
        "API_KEY": "your-api-key"
      }
    }
  }
}

​
Transport types

​
stdio

Spawns the MCP server as a child process that communicates over stdin/stdout. 
{
  "mcpServers": {
    "my-server": {
      "command": "npx",
      "args": ["-y", "some-mcp-server"],
      "env": {}
    }
  }
}
FieldDescription
commandThe command to run
argsArray of command arguments
envEnvironment variables for the process

​
HTTP

Connects to a remote MCP server over HTTP. 
{
  "mcpServers": {
    "remote-server": {
      "url": "https://mcp.example.com/sse"
    }
  }
}
FieldDescription
urlThe server endpoint URL

​
Managing MCP servers

Click MCP in the sidebar to open the MCP panel. The panel shows:
  • A list of configured servers with their connection status
  • Green dot — connected, with the number of available tools
  • Red dot — disconnected or errored
  • An expandable tool list for each server

​
Edit configuration

The MCP panel includes a config editor where you can edit your .mcp.json directly. Click Save to apply changes and Reconnect to re-establish server connections.

​
Tool integration

Once connected, MCP tools appear alongside the built-in tools. The agent can discover and call them like any other tool. Tool results support both text and image content. MCP server connections are cached per project and cleaned up automatically when the app quits.